Update formatter.

Author: liuliu

WORKSPACE

@@ -86,41 +86,41 @@ load("@rules_python//python:pip.bzl", "pip_install")
 new_git_repository(
  name = "SwiftArgumentParser",
  build_file = "swift-argument-parser.BUILD",
- commit = "82905286cc3f0fa8adc4674bf49437cab65a8373",
+ commit = "9f39744e025c7d377987f30b03770805dcb0bcd1",
  remote = "https://github.com/apple/swift-argument-parser.git",
- shallow_since = "1647436700 -0500",
+ shallow_since = "1661571047 -0500",
 )
 
 new_git_repository(
  name = "SwiftSystem",
  build_file = "swift-system.BUILD",
- commit = "836bc4557b74fe6d2660218d56e3ce96aff76574",
+ commit = "025bcb1165deab2e20d4eaba79967ce73013f496",
  remote = "https://github.com/apple/swift-system.git",
- shallow_since = "1638472952 -0800",
+ shallow_since = "1654977448 -0700",
 )
 
 new_git_repository(
  name = "SwiftToolsSupportCore",
  build_file = "swift-tools-support-core.BUILD",
- commit = "b7667f3e266af621e5cc9c77e74cacd8e8c00cb4",
+ commit = "4f07be3dc201f6e2ee85b6942d0c220a16926811",
  remote = "https://github.com/apple/swift-tools-support-core.git",
- shallow_since = "1643831290 -0800",
+ shallow_since = "1659981427 -0700",
 )
 
 new_git_repository(
  name = "SwiftSyntax",
  build_file = "swift-syntax.BUILD",
- commit = "0b6c22b97f8e9320bca62e82cdbee601cf37ad3f",
+ commit = "72d3da66b085c2299dd287c2be3b92b5ebd226de",
  remote = "https://github.com/apple/swift-syntax.git",
- shallow_since = "1647591231 +0100",
+ shallow_since = "1664965455 +0200",
 )
 
 new_git_repository(
  name = "SwiftFormat",
  build_file = "swift-format.BUILD",
- commit = "e6b8c60c7671066d229e30efa1e31acf57be412e",
+ commit = "5f184220d032a019a63df457cdea4b9c8241e911",
  remote = "https://github.com/apple/swift-format.git",
- shallow_since = "1647972246 -0700",
+ shallow_since = "1665415355 -0700",
 )
 
 load("@s4nnc//:deps.bzl", "s4nnc_extra_deps")

examples/t2i-adapter/main.swift

@@ -49,7 +49,8 @@ func Adapter(channels: [Int], numRepeat: Int) -> ((PythonObject) -> Void, Model)
  var outs = [Model.IO]()
  for (i, channel) in channels.enumerated() {
  for j in 0..<numRepeat {
- let (skipModel, inLayerConv2d, outLayerConv2d, resnetBlock) = ResnetBlock(outChannels: channel, inConv: previousChannel != channel)
+ let (skipModel, inLayerConv2d, outLayerConv2d, resnetBlock) = ResnetBlock(
+ outChannels: channel, inConv: previousChannel != channel)
  previousChannel = channel
  out = resnetBlock(out)
  let reader: (PythonObject) -> Void = { state_dict in
@@ -107,9 +108,12 @@ func ResnetBlockLight(outChannels: Int) -> (
  )
 }
 
-func Extractor(prefix: String, channel: Int, innerChannel: Int, numRepeat: Int, downsample: Bool) -> ((PythonObject) -> Void, Model) {
+func Extractor(prefix: String, channel: Int, innerChannel: Int, numRepeat: Int, downsample: Bool)
+ -> ((PythonObject) -> Void, Model)
+{
  let x = Input()
- let inConv = Convolution(groups: 1, filters: innerChannel, filterSize: [1, 1], hint: Hint(stride: [1, 1]))
+ let inConv = Convolution(
+ groups: 1, filters: innerChannel, filterSize: [1, 1], hint: Hint(stride: [1, 1]))
  var out = inConv(x)
  var readers = [(PythonObject) -> Void]()
  for i in 0..<numRepeat {
@@ -127,7 +131,8 @@ func Extractor(prefix: String, channel: Int, innerChannel: Int, numRepeat: Int,
  }
  readers.append(reader)
  }
- let outConv = Convolution(groups: 1, filters: channel, filterSize: [1, 1], hint: Hint(stride: [1, 1]))
+ let outConv = Convolution(
+ groups: 1, filters: channel, filterSize: [1, 1], hint: Hint(stride: [1, 1]))
  out = outConv(out)
  if downsample {
  let downsample = AveragePool(filterSize: [2, 2], hint: Hint(stride: [2, 2]))
@@ -155,7 +160,9 @@ func AdapterLight(channels: [Int], numRepeat: Int) -> ((PythonObject) -> Void, M
  var out: Model.IO = x
  var outs = [Model.IO]()
  for (i, channel) in channels.enumerated() {
- let (reader, extractor) = Extractor(prefix: "\(i)", channel: channel, innerChannel: channel / 4, numRepeat: numRepeat, downsample: i != 0)
+ let (reader, extractor) = Extractor(
+ prefix: "\(i)", channel: channel, innerChannel: channel / 4, numRepeat: numRepeat,
+ downsample: i != 0)
  out = extractor(out)
  outs.append(out)
  readers.append(reader)
@@ -246,14 +253,16 @@ func CLIPResidualAttentionBlock(prefix: String, k: Int, h: Int, b: Int, t: Int)
  return (reader, Model([x], [out]))
 }
 
-func StyleAdapter(width: Int, outputDim: Int, layers: Int, heads: Int, tokens: Int, batchSize: Int) -> ((PythonObject) -> Void, Model)
+func StyleAdapter(width: Int, outputDim: Int, layers: Int, heads: Int, tokens: Int, batchSize: Int)
+ -> ((PythonObject) -> Void, Model)
 {
  let x = Input()
  let lnPre = LayerNorm(epsilon: 1e-5, axis: [2])
  var out = lnPre(x)
  var readers = [(PythonObject) -> Void]()
  for i in 0..<layers {
- let (reader, block) = CLIPResidualAttentionBlock(prefix: "transformer_layes.\(i)", k: width / heads, h: heads, b: batchSize, t: 257 + tokens)
+ let (reader, block) = CLIPResidualAttentionBlock(
+ prefix: "transformer_layes.\(i)", k: width / heads, h: heads, b: batchSize, t: 257 + tokens)
  out = block(out.reshaped([batchSize, 257 + tokens, width]))
  readers.append(reader)
  }
@@ -295,21 +304,26 @@ let hint = torch.randn([2, 3, 512, 512])
 // let adapter = ldm_modules_encoders_adapter.Adapter(cin: 64, channels: [320, 640, 1280, 1280], nums_rb: 2, ksize: 1, sk: true, use_conv: false).to(torch.device("cpu"))
 // let adapterLight = ldm_modules_encoders_adapter.Adapter_light(cin: 64 * 3, channels: [320, 640, 1280, 1280], nums_rb: 4).to(torch.device("cpu"))
 let style = torch.randn([1, 257, 1024])
-let styleAdapter = ldm_modules_encoders_adapter.StyleAdapter(width: 1024, context_dim: 768, num_head: 8, n_layes: 3, num_token: 8).to(torch.device("cpu"))
-styleAdapter.load_state_dict(torch.load("/home/liu/workspace/T2I-Adapter/models/t2iadapter_style_sd14v1.pth"))
+let styleAdapter = ldm_modules_encoders_adapter.StyleAdapter(
+ width: 1024, context_dim: 768, num_head: 8, n_layes: 3, num_token: 8
+).to(torch.device("cpu"))
+styleAdapter.load_state_dict(
+ torch.load("/home/liu/workspace/T2I-Adapter/models/t2iadapter_style_sd14v1.pth"))
 let state_dict = styleAdapter.state_dict()
 print(state_dict.keys())
 let ret = styleAdapter(style)
 print(ret)
 
-let styleEmbed = try Tensor<Float>(numpy: state_dict["style_embedding"].type(torch.float).cpu().numpy())
+let styleEmbed = try Tensor<Float>(
+ numpy: state_dict["style_embedding"].type(torch.float).cpu().numpy())
 
 let graph = DynamicGraph()
 let hintTensor = graph.variable(try! Tensor<Float>(numpy: hint.numpy())).toGPU(0)
 let styleTensor = graph.variable(try! Tensor<Float>(numpy: style.numpy())).toGPU(0)
 // let (reader, adapternet) = Adapter(channels: [320, 640, 1280, 1280], numRepeat: 2)
 // let (reader, adapternet) = AdapterLight(channels: [320, 640, 1280, 1280], numRepeat: 4)
-let (reader, styleadapternet) = StyleAdapter(width: 1024, outputDim: 768, layers: 3, heads: 8, tokens: 8, batchSize: 1)
+let (reader, styleadapternet) = StyleAdapter(
+ width: 1024, outputDim: 768, layers: 3, heads: 8, tokens: 8, batchSize: 1)
 graph.workspaceSize = 1_024 * 1_024 * 1_024
 graph.withNoGrad {
  // let hintIn = hintTensor.reshaped(format: .NCHW, shape: [2, 3, 64, 8, 64, 8]).permuted(0, 1, 3, 5, 2, 4).copied().reshaped(.NCHW(2, 64 * 3, 64, 64))